JP2005184013A - Method for measuring film thickness of semiconductor layer and method for manufacturing semiconductor substrate - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To accurately measure a film thickness by considering a node in which an interference of a time lag to acquire interference information with a light is weakened. <P>SOLUTION: A method for measuring the film thickness of an active layer 6a for an SOI substrate 6 constituted by interposing an oxide film 6c between the active layer 6a and a support substrate 6b comprises the step of first setting an analyzing wavelength range so as to remove a wavelength to become the 'node' in which the interference is weakened by reflected lights on both surfaces of the film 6b. The method further comprises the steps of dispersing the reflected light by irradiating the layer 6a with the light by a diffraction grating 4a according to respective wavelengths, and then simultaneously acquiring interference information of the light according to the dispersed wavelengths by a CCD array 4b. The method also comprises the step of thereafter calculating the film thickness of the layer 6a by using the interference information in the analyzing wavelength range. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a method for measuring the thickness of an active layer (semiconductor layer) in an SOI substrate in which an oxide film is sandwiched between two silicon substrates, and a method for manufacturing a semiconductor substrate such as an SOI substrate.

  The silicon wafer is subjected to a lapping process and an etching process after a slicing process, and then a polishing process is performed and mirror polishing is performed. Specifically, this polishing is performed by attaching a plurality of wafers to a plurality of rotating polishing blocks with wax or the like, and then pressing the wafer against the surface of the rotating polishing platen to which the rotating polishing platen to which the polishing cloth is bonded is bonded. It is done by that.

  In recent years, demands for flatness and parallelism of silicon wafers have become strict, and in order to satisfy these demands, it is necessary to accurately control the thickness of the silicon wafer. In particular, when an SOI substrate is formed by bonding two silicon substrates, it is important to control the film thickness during polishing in order to obtain an active layer having a predetermined thickness.

  However, in the conventional SOI substrate polishing, each step of grinding and rough polishing is performed for a predetermined time, and then the SOI substrate is taken out of the polishing apparatus, and then the thickness of the active layer is measured. Adjustment polishing and film thickness measurement were repeated until the thickness was reached. In other words, if excessive polishing is performed, the SOI substrate becomes unusable, so adjustment polishing and film thickness measurement must be repeated, and the film thickness measurement must be performed by removing it from the polishing apparatus each time. I had to. For this reason, there have been problems such as an increase in the number of work steps and an increase in work time for polishing the SOI substrate.

  On the other hand, various techniques for monitoring and monitoring the film thickness during polishing have been studied. And as the most promising method, there is a film thickness measuring method using interference of light (see Patent Document 1: Japanese Patent Laid-Open No. 8-21616).

  However, as a result of repeated investigations and examinations by the inventors, even in film thickness measurement using light interference, a time lag occurs while acquiring interference information for each wavelength. It has been found that there is a problem that the film thickness changes and accurate film thickness measurement cannot be performed.

  In addition, in an SOI substrate configuration in which an oxide film is sandwiched between two silicon substrates, the oxide film is sandwiched between materials having the same optical characteristics, so that light caused by interference is reflected by reflected light on both sides of the oxide film. Canceled, a “node” portion where light interference (light reflectance) is weakened according to the oxide film thickness is generated. For this reason, it has been found that there is a problem that measurement becomes difficult in the wavelength region in the vicinity of the “node”.

  For example, when the thickness of the active layer is 5 μm and the thickness of the oxide film is 1 μm and 2 μm, the relationship between the wavelength and the reflectance when the film thickness simulation is performed is shown in FIGS. Show. As shown in FIG. 23A, the light interference waveform has a maximum value in the vicinity where the wavelength is 830 nm. If there is a “node” in the vicinity of the wavelength, as shown in FIG. The position becomes the minimum value. Moreover, it may not become a peak, and it may become a maximum value. Thus, if an accurate peak cannot be obtained at a position where it should be the original peak, accurate film thickness measurement cannot be performed.

In view of the above points, an object of the present invention is to enable accurate film thickness measurement in consideration of a time lag for acquiring interference information. It is another object of the present invention to accurately measure a film thickness in consideration of a node where light interference is weakened.
JP-A-8-216061

  In order to achieve the above object, according to the first aspect of the present invention, an SOI substrate (6) in which an oxide film (6c) is sandwiched between an active layer (6a) composed of a semiconductor layer and a support substrate (6b). Is a film thickness measuring method that measures the film thickness of the active layer based on the reflected light from the SOI substrate, and the interference of light is weakened by the reflected light on both sides of the oxide film. The step of setting the analysis wavelength region so as to exclude the reflected wavelength, the step of splitting the reflected light by irradiating the active layer with respect to each wavelength, and the interference information of the divided light for each wavelength. And a step of acquiring all at once, and a step of calculating the thickness of the active layer using interference information in the analysis wavelength region.

  In this way, the interference information is acquired in a lump so that there is no time lag, and the analysis wavelength region is set so as to exclude the wavelength at which the light interference is weakened, that is, the “nodes”. Based on the peak value, the thickness of the active layer can be measured. For this reason, it is possible to accurately measure the thickness of the active layer. For example, as shown in claim 3, in the step of setting the analysis wavelength region, assuming that the refractive index of the oxide film is nox, the film thickness is dox, and m is an arbitrary positive number, It sets as (lambda) which shows an analysis wavelength area | region.

  Also, in the step of calculating the thickness of the active layer, it is possible to calculate the thickness of the active layer from the wavelength that is the maximum or minimum value of the waveform indicating the relationship between the wavelength obtained from the interference information and the reflection intensity. is there. According to a fourth aspect of the present invention, if the film thickness of the active layer is calculated from the wavelength that is the minimum value among the wavelengths that are the maximum value or the minimum value, the wavelength that is less affected by noise and has a peak value. Can be detected more accurately.

In claim 1, the wavelength at an arbitrary maximum or minimum value is λ ₁ , λ ₂ , the refractive index of the film at the wavelength λ is n (λ), the frequency or wave number between both wavelengths is X, and the single layer film structure And the thickness d of the active layer is calculated by the above equation ( ₁₎ , where s ₁ and s ₂ are the phase shift amounts at the maximum or minimum wavelength in the two-layer film structure.

  By calculating the film thickness of the active layer using such a calculation formula, it is possible to correct an error in the measured film thickness that may occur due to the relationship between the film thickness of the active layer and the oxide film. Measurement can be made possible.

The phase shift amounts s ₁ and s ₂ are obtained on the basis of the phase value of the oxide film obtained from the thickness of the oxide film and the wavelength of light passing through the oxide film. .

  The invention according to claim 5 is characterized in that, in the step of setting the analysis wavelength region, a wavelength at which the light transmittance in the active layer is a predetermined value or more is set as the analysis wavelength region. For example, as shown in claim 6, the wavelength at which the light transmittance in the active layer is 10% or more is set as the analysis wavelength region. In this way, the analysis wavelength region can be set in a region where light interference can be clearly confirmed.

  In the invention according to claim 7, in the step of separating the reflected light and the step of obtaining the interference information collectively, the diffraction grating (4a) for separating the reflected light and each wavelength of the light dispersed by the diffraction grating A spectroscope (4) provided with a CCD array (4b) that collects other interference information at once is used. With such a configuration, it is possible to collect reflected light spectrum and interference information collectively.

  When a CCD array is used, as shown in claim 8, it is preferable to set a wavelength of 1000 nm or less as the analysis wavelength region. That is, when the wavelength is 1000 nm or more, there is a possibility that the relationship between the wavelength and the reflection intensity cannot be accurately obtained due to the sensitivity reduction of the CCD array. For this reason, the wavelength which becomes a peak value correctly can be calculated | required by carrying out like this claim.

  The invention according to claim 9 has a polishing step of polishing the active layer with respect to the SOI substrate, and the thickness of the active layer during the polishing is the semiconductor layer according to any one of claims 1 to 8 The method is characterized in that the polishing is terminated when it is detected that the active layer has reached a desired film thickness.

  As described above, by using the semiconductor layer thickness measuring method according to any one of claims 1 to 8 for an SOI substrate in which the thickness adjustment of the active layer is performed by polishing, a polishing apparatus can be used. The thickness of the active layer can be accurately measured without taking out the SOI substrate.

  Such film thickness measurement may be performed with the SOI substrate being rotated for polishing, as shown in claim 10.

  The invention according to claim 11 is the film thickness of the first layer in the substrate (6) having the first layer (6a) and the second layer (6c) having a refractive index different from that of the first layer. The invention according to claim 1 is applied to the film thickness measuring method for measuring the thickness. Thus, the invention according to claim 1 can be applied to the two-layer film structure having different refractive indexes, and the film thickness can be measured with high accuracy.

  In addition, the code | symbol in the bracket | parenthesis of each said means shows the correspondence with the specific means as described in embodiment mentioned later.

(First embodiment)
A method for measuring the thickness of the semiconductor layer in the first embodiment of the present invention will be described. FIG. 1 shows a schematic configuration of a film thickness measuring apparatus that realizes the film thickness measuring method in the present embodiment. Hereinafter, the configuration of the film thickness measuring apparatus will be described with reference to FIG.

  As shown in FIG. 1, the film thickness measurement apparatus includes a light source 1, an optical fiber 2, a measurement head 3, a spectrometer 4, and an analysis device 5. As the light source 1, for example, a tungsten halogen lamp that generates white light is used. As the light source 1, light obtained by separating a part of white light generated by the tungsten halogen lamp with a filter or the like may be used, and an incandescent bulb or a xenon lamp may be used. The optical fiber 2 transmits light composed of, for example, a plurality of fibers, and has a role of transmitting light from the light source 1 to the measurement head and a role of transmitting light from the measurement head 3 to the spectrometer 4. Plays.

  The measurement head 3 plays the role of irradiating the light transmitted from the optical fiber 2 toward the surface of the active layer 6 a of the SOI substrate 6 and receiving the reflected light from the SOI substrate 6. The reflected light from the SOI substrate 6 received from the measurement head 3 in this way is sent to the spectroscope 4 through the optical fiber 2.

  The spectroscope 4 divides the transmitted light into components, and the spectroscope 4 can receive light interference information all at once. FIG. 2 shows an example of a specific configuration of the spectrometer 4. As shown in this figure, the spectroscope 4 is provided with a diffraction grating 4a and a CCD array 4b. If the reflected light from the SOI substrate 6 sent through the optical fiber 2 is detected light, it is diffracted. The configuration is such that the detection light is split for each wavelength by the grating 4a, and the split light for each wavelength is collectively received by the CCD array 4b as interference information. For this reason, the interference information of light for each wavelength can be acquired without a time lag. In the present embodiment, the CCD array 4b out of these components corresponds to means for collectively receiving light interference information.

  The analysis device 5 analyzes the film thickness of the active layer 6a on the surface of the SOI substrate 6 on the basis of the light spectrally separated by the spectroscope 4, specifically, interference information taken into the CCD array 4b. Generates output according to the thickness analysis result. By this output, for example, monitoring of the film thickness analysis result by the display device is performed, further film thickness adjustment based on the film thickness analysis result, and detection of the polishing stop point of the active layer 6a are performed.

  Then, the flowchart of the film thickness measurement using the film thickness measuring apparatus shown in FIG. 1 is shown in FIG. 3, and the details of the film thickness measurement will be described based on this figure. First, an analysis wavelength region is set (step S101). This analysis wavelength region means a wavelength region used for analysis of film thickness measurement in the interference information received by the CCD array 4b. Details of the analysis wavelength region setting method will be described later.

  Next, the light source 1 is caused to emit light, and the surface of the active layer 6a in the SOI substrate 6 is irradiated through the optical fiber 2 and the measurement head 3 (step S102), and then the SOI substrate 6 is transmitted through the measurement head 3 and the optical fiber 2. The reflected light from is sent to the spectroscope 4, and the spectroscope 4 receives interference information for each wavelength (step S103). Thereby, interference information for each wavelength is acquired without causing a time lag. Then, in the analysis device 5, the waveform of the reflected light, that is, the relationship between the wavelength and the reflection intensity is obtained from the interference information for each wavelength. When this relationship is illustrated, the reflection intensity periodically changes with respect to the wavelength. Therefore, based on the relationship between the wavelength and the reflection intensity, the peak value (maximum value and minimum value) in the waveform of the reflected light and the wave number in the analysis wavelength region set above are obtained (step S104). Thereafter, the film thickness is calculated by the analysis device 5 based on the obtained peak value and wave number (step S105), and an output corresponding to the film thickness analysis result is performed (step S106).

  Here, a film thickness analysis method and an analysis wavelength region setting method will be described. First, the film thickness analysis method will be described with reference to the reference diagram shown in FIG.

  FIG. 4 illustrates the relationship between incident light and reflected light when the SOI substrate 6 is irradiated with light. As shown in this figure, since the structure of the SOI substrate 6 is such that the oxide film 6c is sandwiched between the photoactive layer 6a and the support substrate 6b, when the light is irradiated, the active layer 6a and the oxide film 6c Reflection occurs at the interface or the interface between the oxide film 6c and the support substrate 6b, and these reflected lights interfere with each other. This light interference varies depending on the film thickness d, the refractive index n, and the wavelength λ of the active layer 6a.

  When the refractive index of the outside air is no, the film thickness of the oxide film 6c is dox, the refractive index is nox, and the refractive index of the support substrate 6b is ns, the peak value of the reflectance becomes the following equation. It can be seen which of the equations showing the peak value of the reflectivity shows the maximum value and the minimum value according to the relationship between the refractive indexes n, nox and ns of ˜6c. In the following equation, m is a natural number.

すなわち、ｎｏ＞ｎの時にｎ＞ｎｓであれば数４の第２式が極大値、第１式が極小値となり、ｎ＜ｎｓであれば数４の第１式が極大値、第２式が極小値となる。また、ｎｏ＜ｎの時にｎ＞ｎｓであれば数４の第１式が極大値、第２式が極小値となり、ｎ＜ｎｓであれば数４の第２式が極大値、第１式が極小値となる。

That is, if no> n and n> ns, then the second equation of Equation 4 is the maximum value, the first equation is the minimum value, and if n <ns, the first equation of Equation 4 is the maximum value, and the second equation Becomes the minimum value. Further, if n> ns when no <n, the first expression of Formula 4 is the maximum value and the second expression is the minimum value, and if n <ns, the second expression of Formula 4 is the maximum value, and the first expression Becomes the minimum value.

  And since the position where the peak value of the reflectance coincides with the position where the waveform of the reflected light in step S104 described above becomes the peak value, the film of the active layer 6a from the relationship of the first and second formulas of the above equation 4 The thickness d is determined as follows:

ただし、λ₁、λ₂は任意のピーク値における波長、ｎ（λ）は波長λでの膜の屈折率、Ｘは両波長間の周波数（波数）を示している。なお、ここでいうピーク値における波長とは、極大値、極小値いずれのピーク値における波長であっても良い。つまり、極大値同士、極小値同士、または極大値と極小値との組み合わせにおける波長いずれを用いても良い。ただし、極大値と極小値との組み合わせにおける波長を用いる場合、Ｘが自然数ではなく自然数に対して±０．５を加算した値となる。

Here, λ ₁ and λ ₂ are wavelengths at arbitrary peak values, n (λ) is the refractive index of the film at the wavelength λ, and X is the frequency (wave number) between the two wavelengths. Here, the wavelength at the peak value may be the wavelength at the peak value of either the maximum value or the minimum value. In other words, any wavelength may be used between the local maximum values, the local minimum values, or the combination of the local maximum value and the local minimum value. However, when the wavelength in the combination of the maximum value and the minimum value is used, X is not a natural number but a value obtained by adding ± 0.5 to the natural number.

For reference, FIG. 5 shows the results of examining the relationship between the wavelength and the reflection intensity when the thickness of the active layer 6a is 14 μm and the thickness of the oxide film 6c is 1.3 μm. As shown in this figure, assuming that the wavelength at the peak value at the lowest wavelength is λ ₁ , the wavelength at the peak value when X = 6 is λ ₂ .

  Subsequently, the analysis wavelength region will be described. In setting the analysis wavelength region, the wavelength region from which the above-mentioned “nodes” are removed is set as the wavelength used for the analysis. First, prior to the analysis wavelength region setting method, the principle of “nodes” and how “nodes” affect film thickness measurement will be described.

FIG. 6 schematically shows the light interference phenomenon, and the principle of occurrence of “nodes” will be described based on this figure. First, assuming that the active layer 6a, the oxide film 6c, and the support substrate 6b are media I, II, and III, respectively, the refractive indexes in these media are n ₁ , n ₂ , and n ₃ , and the incident light amplitude is 1. The amplitude (amplitude reflectance) r _{1 of the} reflected light and the amplitude (amplitude reflectance) t ₁ of the transmitted light are expressed by the following equations from Fresnel's law.

また、媒質ＩＩ、ＩＩＩの界面での振幅反射率ｒ₂および振幅透過率ｔ₂も同様に次式のように示される。

Similarly, the amplitude reflectance r ₂ and the amplitude transmittance t ₂ at the interface between the media II and III are also expressed by the following equations.

そして、媒質ＩＩの上下の界面で多重反射するため、媒質Ｉへ出てくる光の振幅反射率ｒ、エネルギー反射率Ｒは次式のように示される。

Since multiple reflection is performed at the upper and lower interfaces of the medium II, the amplitude reflectance r and the energy reflectance R of the light coming out of the medium I are expressed by the following equations.

ただし、δ＝４πｎ₂ｄ／λである。そして、媒質Ｉを活性層６ａ（Ｓｉ）、媒質ＩＩを酸化膜６ｃ（ＳｉＯ₂）、媒質ＩＩＩを支持基板６ｂ（Ｓｉ）とした場合、つまり媒質Ｉ、ＩＩ、ＩＩＩの各材質をＳｉ、ＳｉＯ₂、Ｓｉとした場合で考えると、ｎ₁＝ｎ₃であるため、数１０が求められ、この数１０と上記数８とにより、数１１が求められる。

However, δ = 4πn ₂ d / λ. When the medium I is the active layer 6a (Si), the medium II is the oxide film 6c (SiO ₂ ), and the medium III is the support substrate 6b (Si), that is, the materials of the mediums I, II, and III are Si, SiO. Considering the case of ₂ and Si, since n ₁ = n ₃ , Equation 10 is obtained, and Equation 11 is obtained from Equation 10 and Equation 8 above.

そして、ｃｏｓδは−１〜１の間で変化するため、ｃｏｓδ＝１となるとき、Ｒは極小値０をとる。すなわち、酸化膜６ｃの界面での反射が実質的に起こらなくなる。従って、４πｎ₂ｄ／λ＝２ｈπ（ｈは自然数）の条件を満たす時に「節」が現われる。このため、この「節」を避けるように、解析波長領域の設定を行う。図７を用いて、この解析波長領域の設定方法を説明する。

Since cos δ changes between −1 and 1, when cos δ = 1, R takes the minimum value 0. That is, reflection at the interface of the oxide film 6c substantially does not occur. Therefore, a “node” appears when the condition of 4πn ₂ d / λ = 2hπ (h is a natural number) is satisfied. Therefore, the analysis wavelength region is set so as to avoid this “section”. The analysis wavelength region setting method will be described with reference to FIG.

  FIG. 7 shows the result of examining the relationship between the wavelength and the reflection intensity when the thickness of the active layer 6a is 14 μm and the thickness of the oxide film 6c is 1.3 μm. In this figure, “nodes” are generated in the vicinity of the wavelength of 957 nm, and the maximum value and the minimum value that should be generated in the vicinity of this wavelength originally do not appear clearly.

  Therefore, if this “node” portion is included in the measurement range, it becomes difficult to detect the peak value and correct film thickness measurement cannot be performed. Further, in order to surely detect the maximum value and the minimum value, it is desirable that there is a change amount of about ± 10% or more, and using the measurement result of FIG. 7, ± 5 nm from the center wavelength of the “node”. It is preferable to use the wavelength in the range excluding the degree as the analysis wavelength region. Since the wavelength dependency of the amplitude amount (reflectance change amount) due to the thickness of the active layer 6a varies depending on the wavelength and the thickness of the oxide film 6c, the “node” shown in the measurement result of FIG. In general, the wavelength ± 5 nm is expressed by the following equation.

なお、本実施形態のようにＳＯＩ基板６を用いる場合には、屈折率ｎｏｘや膜厚ｄｏｘが予め分かっていることから、これらの値を予め入力しておくことによって、数１２から「節」を求めることができる。図７の解析結果においては、数１２における酸化膜６ｃの屈折率ｎｏｘが１．４５、膜厚ｄｏｘが１．３２μｍ、自然数ｍが４となる。従って、解析波長領域として好ましい範囲は、次式のように示される。

Note that when the SOI substrate 6 is used as in the present embodiment, the refractive index nox and the film thickness dox are known in advance. Can be requested. In the analysis result of FIG. 7, the refractive index nox of the oxide film 6 c in Expression 12 is 1.45, the film thickness dox is 1.32 μm, and the natural number m is 4. Therefore, a preferable range as the analysis wavelength region is expressed by the following equation.

参考として、活性層６ａの膜厚を９μｍ、酸化膜６ｃの膜厚を３．５４μｍとした場合における解析結果を図８に示す。この図に示されるように、「節」の中心波長から離れると急激に振幅が大きくなっている。この場合、±１０％程度以上の変化量を見込むと、「節」から約±２ｎｍほど外側の波長を解析波長領域とするのが好ましいと言える。このように、酸化膜６ｃの膜厚が厚くなるほど、解析波長領域から除去するべき範囲が狭くなり、上記数１３で示される範囲が妥当であることが分かる。

As a reference, FIG. 8 shows an analysis result when the thickness of the active layer 6a is 9 μm and the thickness of the oxide film 6c is 3.54 μm. As shown in this figure, the amplitude suddenly increases as the distance from the center wavelength of the “node” increases. In this case, when the amount of change of about ± 10% or more is expected, it can be said that it is preferable to set the wavelength outside from the “node” by about ± 2 nm as the analysis wavelength region. Thus, it can be seen that as the thickness of the oxide film 6c increases, the range to be removed from the analysis wavelength region becomes narrower, and the range represented by Equation 13 is appropriate.

  Next, by the method described above, the exposure time in the CCD array 4b is 10 ms, the measurement wavelength interval is 0.3 nm, the spectral resolution (half width) of the spectrometer 4 is 1 nm, and the thickness of the active layer 6a of the SOI substrate 6 is 3 μm. The film thickness was measured using a sample having a thickness of 1.1 μm of the oxide film 6c.

  In the case of such a sample, the analysis wavelength region obtained from Equation 13 is 647 to 802 nm, or 810 to 1068 nm. When the relationship between the wavelength of reflected light and the reflection intensity in the case of using the above sample is illustrated, the relationship with the analysis wavelength region is shown as in FIGS. However, FIG. 9 shows the case where the local minimum values are used as the peak values for film thickness measurement, and FIG. 10 shows the case where the local maximum values are used as the peak values for film thickness measurement.

  As can be seen from these figures, since the analysis wavelength region is set so as to avoid “nodes”, the peak value that is the wave number, maximum value, and minimum value can be accurately measured in this region. . For this reason, it becomes possible to accurately measure the film thickness of the active layer 6a based on the accurate wave number and peak value. Specifically, as shown in the upper right in the figure, the film thickness measurement result of the active layer 6a is the film thickness d = 3.17 μm in the case of FIG. 9, and the film thickness d = 3.20 μm in the case of FIG. It can be seen that the film thickness can be measured accurately.

  As described above, in this embodiment, the interference information of light for each wavelength can be acquired collectively, and the analysis wavelength region is set so as to avoid “section”. It becomes possible to measure the film thickness of the active layer 6a based on the peak value. For this reason, the film thickness of the active layer 6a can be accurately measured.

  In addition, here, as a peak value used for film thickness measurement, it is shown as a case where local minimum values are used (FIG. 9) and a case where local maximum values are used (FIG. 10). In the case of the combination, it is as shown in FIG. As can be seen from this figure, the wave number and the peak value can be accurately obtained even with the combination of the maximum value and the minimum value, and the active layer 6a can be accurately determined. The film thickness measurement result in this case is film thickness d = 3.18 μm, and it can be seen that accurate film thickness measurement can also be performed in this case.

  As described above, the peak value used for film thickness measurement may be either a maximum value or a minimum value. However, since the waveform of the minimum value tends to be sharper than the maximum value, the minimum value is preferably used. Better to be. This is related to accurately detecting the position where the maximum value or the minimum value is obtained. That is, as the film thickness of the active layer 6a is reduced, the change in reflectance near the maximum value due to the wavelength becomes gradual, so that it becomes difficult to detect the peak value due to the influence of noise. To cope with this, it is possible to take measures such as estimating an ideal line by averaging during data processing, but this is not preferable because the processing becomes complicated. On the other hand, since the change in reflectance near the minimum value due to the wavelength is abrupt, it is difficult to be affected by noise, and the wavelength at the peak value can be detected more accurately. For this reason, it is preferable to use a minimum value as a peak value used for film thickness measurement.

  For reference, FIGS. 12 to 14 show the relationship between the wavelength of reflected light and the reflection intensity when the thickness of the active layer 6a is measured by appropriately changing the thickness of the active layer 6a and the oxide film 6c. .

  FIG. 12 shows a case where the thickness of the active layer 6a is set to be as large as 25 μm and the thickness of the oxide film 6c is set to 1.1 μm. Even in such a case, by setting the analysis wavelength region, the wave number and the peak value can be accurately obtained, and the active layer 6a can be accurately obtained. However, when the active layer 6a is set to be thicker in this way, the peak value does not appear in the region having a wavelength shorter than 800 nm due to light absorption in the active layer 6a. For this reason, it is necessary to set the analysis wavelength region in consideration of light absorption. For example, FIG. 13 shows a case where the active layer 6a is made thicker (45 μm) than FIG. 12, but it can also be seen from this figure that light absorption by the active layer 6a increases according to the thickness. .

  FIG. 15 shows the dependence of the wavelength on the light absorption coefficient α of the active layer 6a, that is, Si. As shown in this figure, the larger the wavelength, the smaller the absorption coefficient α and the better the light transmission. Then, when the relationship between the light interference and the transmittance is examined from the transmittance at the active layer 6a and the results shown in FIGS. 12 and 13, the light interference can be clearly confirmed when the transmittance reaches 10%. I understood. FIG. 16 shows the relationship between the thickness of the active layer 6a at which the transmittance is 10% and the wavelength. Interference of light can be clearly confirmed in the shaded area in this figure, and the film thickness can be measured. That is, when the thickness of the active layer 6a is 25 μm (that is, FIG. 12), light interference is clearly recognized from a wavelength near 820 nm, and the thickness of the active layer 6a is 45 μm (that is, FIG. 13). Then, since the interference of light is clearly recognized from a wavelength near 880 nm, it agrees with the result shown in FIG. For this reason, by setting the analysis wavelength region so that the light transmittance in the active layer 6a is 10% or more, light interference can be obtained accurately and accurate film thickness measurement can be performed. It becomes.

  On the other hand, FIG. 14 shows a case where the active layer 6a is 14 μm and the oxide film 6c is 1.3 μm. Even in this case, the “section” appears clearly as shown in the figure, but since the analysis wavelength region is set to avoid this “section”, the wave number and peak value can be obtained accurately. Thus, the active layer 6a can be accurately obtained.

  Moreover, when the measurement data etc. in the case of using the said method and the film thickness measurement data using a standard measuring device were compared, it represented as FIG. However, the measurement conditions when the measurement data is obtained are the exposure time 6 msec in the CCD array 4b, the measurement distance 20mm from the measurement head 3 to the SOI substrate 6, and the measurement data is obtained by detecting the maximum wavelength. ing. When the thickness of the oxide film 6c is 1.1 μm, it is 850 to 960 nm (however, only when the thickness of the active layer 6a is 46 μm is 920 to 1000 nm), when it is 1.3 μm, it is 820 to 920 nm, and when it is 3.5 μm, it is 860. An analysis wavelength region of ˜920 nm is set. The standard measuring instrument uses Nanometrics nanometrics when the thickness of the active layer 6a is 7 μm or less, and FT-IR when the thickness is 7 μm or more.

  As can be seen from the measurement data, each measurement data almost coincides with the film thickness measurement data of the standard measuring instrument indicated by the solid line, and it can be seen that a very good correlation is obtained.

  In addition, measurement was performed when the SOI substrate 6 was rotated using the above method. The result is shown in FIG. In this measurement, the measurement conditions are set so that the exposure time in the CCD array 4b is 10 milliseconds, the analysis wavelength region is 850 to 960 nm, and the rotation speed of the SOI substrate 6 is 75 rotations / minute. Then, an SOI substrate 6 having an active layer 6a thickness of 9.3 μm, an oxide film 9c thickness of 1.1 μm, and a diameter of 6 inches is used, and the measurement interval is 10 mm from the outermost periphery of the SOI substrate 6. Measurements are taken every 1 second of time interval. In addition, the broken line in a figure has shown the maximum value and minimum value when the film thickness of the active layer 6 is measured in four places in the state in which the SOI substrate 6 was stationary using the FT-IR apparatus.

  As can be seen from this result, even when the SOI substrate 6 is rotated, the film thickness of the active layer 6a is accurately measured in accordance with the measurement result of the FT-IR apparatus used as a standard measuring instrument. Can be said.

  Further, the above method was used to measure the case where water was interposed on the surface of the SOI substrate 6 (active layer 6a). The result is shown in FIG. In this measurement, the thickness of water is 1 mm, the thickness of the active layer 6a is 14 μm, and the thickness of the oxide film 6c is 1.3 μm. In this way, even when water is present, by selecting the analysis waveform region so as to avoid “nodes”, the wave number and peak value can be accurately obtained, and the active layer 6a can be accurately obtained. It becomes.

  In the above-described setting of the analysis wavelength region, the upper limit of the analysis wavelength region has not been described, but the upper limit of the analysis wavelength region is set depending on what is used as a means for collectively receiving reflected light interference information. You can also. For example, as used in the description herein, when the CCD array 4b is used as a means for collectively receiving reflected light interference information, the actually measured waveform of the wavelength and the reflection intensity is as shown in FIG. In the region where the wavelength is long, it differs from the theoretical waveform shown in FIG. This occurs due to insufficient sensitivity of the CCD array 4b. When the CCD array 4b is used, insufficient sensitivity occurs from around the wavelength of about 1000 nm. For this reason, when using the CCD array 4b, it is preferable to set the upper limit of the analysis wavelength region to about 1000 nm.

(Second Embodiment)
A second embodiment of the present invention will be described. In the present embodiment, the film thickness measurement apparatus shown in FIG. 1 is used to measure the film thickness of the active layer 6a during polishing of the SOI substrate 6 by the polishing apparatus. FIG. 21 is a schematic diagram showing the state of film thickness measurement in the present embodiment.

  The polishing apparatus shown in FIG. 21 is a CMP (Chemical mechanical polish) apparatus. This polishing apparatus includes a polishing pad 11, a surface plate 12 on which the polishing pad 11 is stuck, a head 13 to which an SOI substrate 6 is attached, and a slurry supply unit 14 for dropping slurry 14a containing abrasive grains. .

  Then, after the SOI substrate 6 is fixed to the head 13 so that the planarized surface (that is, the active layer 6a side) of the SOI substrate 6 is located on the polishing pad 11 side, the slurry 14a is supplied from the slurry supply unit 14. The surface of the SOI substrate 6 can be flattened by the polishing pad 11 by rotating the surface plate 12 and the head 13 with the head 13 pressing the SOI substrate 6 against the surface plate 12.

  In the polishing apparatus configured as described above, the window 12a is provided in the surface plate 12 and the polishing pad 12, and the film thickness is measured through the window 12a. Note that a film thickness measurement apparatus used for film thickness measurement has the same configuration as that shown in FIG.

  The film thickness measurement during such polishing is performed according to the flowchart shown in FIG. That is, first, analysis wavelength region setting, light irradiation to the SOI substrate 6, measurement of reflected light signal by the spectroscope 4, detection of peak position and wave number, film thickness calculation, film thickness output (steps S201 to 206) are performed. . These processes are performed in the same manner as steps S101 to S106 in FIG. Subsequently, it is determined whether or not the measured film thickness is equal to or less than a desired set film thickness (step S207). Thereby, it is determined whether or not the active layer 6a has a desired film thickness. If the determination here is negative, the previous processing is repeated. If the determination is positive, an end point signal indicating the end of polishing is output from the analysis device 5 (step S208). Polishing at is stopped.

  Thus, the film thickness of the active layer 6a can be measured even during the polishing process by the polishing apparatus, and the active layer 6a can be more accurately set to a desired film thickness.

  Here, the film thickness is measured during polishing by providing the window 12 a on the surface plate 12. However, even if the window 12 is not provided on the surface plate 12, a part of the SOI substrate 6 is part of the surface plate 12. Further, the thickness of the active layer 6a may be measured by projecting outside the outer peripheral edge and irradiating the projected portion with light. Further, when a part of the SOI substrate 6 is protruded from the outer peripheral end of the surface plate 12 in this way, the polishing may be performed or the polishing may be temporarily stopped.

(Third embodiment)
In the present embodiment, a case will be described in which the shift amount of the film thickness of the active layer 6a shown in the first embodiment is corrected.

  In each of the above embodiments, since the SOI substrate 6 has a two-layer film structure, it is affected by the oxide film 6c immediately below the active layer 6a, and the phase at the peak wavelength is a single-layer film. Will shift slightly. For this reason, an error occurs in the measured film thickness value, specifically, the film thickness is detected to be thicker than the actual film thickness. May be unacceptable. To solve this problem, it is possible to create a correspondence table from many measured measurement values and perform correction, but it is not versatile and it is not based on the measurement principle, so the variation in correction accuracy will increase. Problems are expected to occur. In particular, the thinner the active layer 6a and the thicker the oxide film 6c, the greater the influence and the greater the measurement error of the film thickness.

  For this reason, in the present embodiment, the amount of shift in the film thickness of the active layer 6a due to the phase shift amount s at the peak wavelength is corrected using the equation shown in equation (14). The calculation formula for correction represented by Equation 14 is obtained by correcting the phase shift amount s at the wavelength corresponding to each peak value from X in Equation 5 shown in the first embodiment.

以下、この活性層６ａの膜厚のズレ量の補正に用いる位相シフト量ｓについて説明するが、それに先立ち、このズレ量の補正の概念に関して分光法の原理を基に説明する。ただし、分光法の原理のうち第１実施形態と同様の部分については、第１実施形態を参照する。

Hereinafter, the phase shift amount s used for correcting the shift amount of the film thickness of the active layer 6a will be described. Prior to that, the concept of correcting the shift amount will be described based on the principle of spectroscopy. However, the first embodiment is referred to for the same part as the first embodiment in the principle of spectroscopy.

  First, in FIG. 6 shown in the first embodiment, it is assumed that the medium II has a single-layer film structure, and light emitted from the medium II to the medium I based on the light interference phenomenon shown in the figure. The energy reflectivity R is obtained. Then, the energy reflectance R is expressed as in the above-described formula 9, and this expression is converted into the formula 15.

そして、δ＝４πｎ₂ｄ／λであり、ｃｏｓδが−１〜１の間で変化することから、ｃｏｓδ＝±１となるときにエネルギー反射率Ｒが極大値もしくは極小値をとる。すなわち、４πｎ₂ｄ／λ＝２ｍπ、及び４πｎ₂ｄ／λ＝（２ｍ−１）π（ｍは自然数）の条件を満たす時に、極大値もしくは極小値をとる。従って、２ｎｄ＝ｍλ、２ｎｄ＝（ｍ−１／２）λの２式が導き出され、いずれの式が極大値、極小値になるかは、各媒質Ｉ、ＩＩ、ＩＩＩの屈折率ｎ₁、ｎ₂、ｎ₃の大小関係、つまりこれらの大小関係に基づき上記数６、数７から求まるｒ₁、ｒ₂によって決まる。

Since δ = 4πn ₂ d / λ and cos δ varies between −1 and 1, the energy reflectivity R takes a maximum value or a minimum value when cos δ = ± 1. That is, when the conditions of 4πn ₂ d / λ = 2mπ and 4πn ₂ d / λ = (2m−1) π (m is a natural number) are satisfied, the maximum value or the minimum value is obtained. Therefore, two formulas of 2nd = mλ and 2nd = (m−1 / 2) λ are derived, and which formula has the maximum value and the minimum value is determined by the refractive index n _{1 of} each medium I, II, III. It is determined by r ₁ and r ₂ obtained from the above equations 6 and 7 based on the magnitude relationship between n ₂ and n ₃ , that is, based on these magnitude relationships.

  Next, FIG. 24 shows a schematic diagram of the light interference phenomenon in the case of the two-layer film structure, and considers the multiple reflection in the case of the two-layer structure.

As shown in FIG. 24, mediums I to IV are assumed, the refractive indexes of these mediums I to IV are n _{1 to} n ₄ , the film thickness of medium II is d ₁ , and the film thickness of medium III is d _2. If the amplitude of the incident light is 1, the amplitude of the reflected light is r ₁ to r ₃ , and the amplitude of the transmitted light is t ₁ to t ₃ , the energy reflectance R is a number by the same method as in the case of the single layer film structure. It is obtained like 16. However, r ₂ ′ in Expression 16 is expressed by Expression 17, and corresponds to the combined reflectance of the amplitude of the reflected light in the medium III and the medium IV. Further, δ ₁ = 4πn ₂ d ₁ / λ and δ ₂ = 4πn ₃ d ₂ / λ.

従って、数１６は、単層膜構造の下側の界面での反射率を媒質ＩＩＩの上下界面での合成反射率ｒ₂’で置き換えたもので表され、ｃｏｓδ₁が−１〜１で変化する間にｒ₂’が一定ならば、単層膜モデルと等価になる。しかしながら、実際には、δ₁だけでなくδ₂も同時に変化するため、ｒ₂’も変化し、極大値もしくは極小値をとるときのδ₁の値にズレが発生する。数１７から明らかなように、ｒ₂’はδ₂に依存するため、δ₁のズレ量はδ₂に依存する。

Therefore, Equation 16 is expressed by replacing the reflectance at the lower interface of the single-layer film structure with the combined reflectance r ₂ ′ at the upper and lower interfaces of the medium III, and cos δ ₁ varies from −1 to _1. If r ₂ ′ is constant during this time, it becomes equivalent to the single layer film model. However, in actuality, not only δ ₁ but also δ ₂ changes at the same time, so that r ₂ ′ also changes, and a deviation occurs in the value of δ ₁ when taking the maximum value or the minimum value. As apparent from Equation 17, since r ₂ ′ depends on δ ₂ , the amount of deviation of δ ₁ depends on δ ₂ .

In other words, when the amount of change of δ ₂ with respect to the change of wavelength λ is sufficiently smaller than the amount of change of δ ₁ , there is no problem if Mathematical Formula 15 is replaced with a single-layer film equation, but δ ₂ and δ ₁ When the difference in change amount is small, the film thickness d _{1 of the} medium II cannot be obtained with high accuracy by approximation with a single layer film. Here, when the difference in the amount of change between δ ₂ and δ ₁ is small, specifically, when the value of δ ₁ / δ ₂ = n ₂ d ₁ / n ₃ d ₂ is small, that is, This shows when the film thickness of the medium II is relatively thin or when the film thickness of the medium III is relatively large.

Based on the above principle, when the difference in change amount between δ ₂ and δ ₁ is small, the film thickness d of the active layer 6a cannot be detected with high accuracy, and the shift amount of the film thickness d is corrected. There is a need.

Therefore, in correcting the amount of deviation of δ ₁ from the film thickness d of the active layer 6a expressed by Equation 5, it is necessary to estimate the amount of deviation of δ ₁ by obtaining the amount of deviation of δ _2. . Then, the phase shift amount s at the wavelength [delta] ₁ is the peak value, since r ₂ 'is dependent on the cos [delta] ₂ is a periodic function, [delta] _2/2 [pi, i.e. [delta] ₂ Effect of oxide film 6c Since it can be represented by the decimal part of the phase value (= 2ndox / λ) of the oxide film 6c obtained from the thickness and the wavelength, the phase shift amount s is uniquely determined by using the decimal part of the phase value as a parameter by obtaining these relations in advance. It is possible to ask for it.

  That is, when the relationship between the decimal part of the phase value and the shift amount s is obtained by simulation, the result shown in FIG. 25 is obtained. Based on this relationship diagram, the phase shift s corresponding to the decimal part of the phase value is calculated. Can be sought.

  In this way, by correcting the shift amount of the film thickness of the active layer 6a in consideration of the phase shift amount s, the film thickness of the active layer 6a can be detected with higher accuracy. As a reference, FIG. 26 shows actual measurement results obtained by examining measurement errors when the correction shown in the present embodiment is performed and when the correction is not performed. In this experiment, the thickness of the oxide film 6c is 3.57 μm. As is clear from this figure, it can be said that the film thickness of the active layer 6a can be detected with higher accuracy by performing the correction shown in the present embodiment.

(Other embodiments)
As the principle of film thickness analysis in each of the above embodiments, the method shown in the first embodiment is not necessarily used, and another film thickness analysis principle may be used. For example, the following method may be used. This method will be described with reference to FIG.

  As shown in FIG. 4, when the SOI substrate 6 is irradiated with light, reflection occurs at the interface between the active layer 6a and the oxide film 6c and the interface between the oxide film 6c and the support substrate 6b, and these reflected lights interfere with each other. . This light interference varies depending on the film thickness d, the refractive index n, and the wavelength λ of the active layer 6a.

When the wavelength changes from λ ₁ to λ ₂ , the reflected light repeatedly intensifies and weakens each other, and assuming that the phase changes by ψ [rad], the thickness d of the active layer 6a is related to interference. It is expressed by the following formula.

このψがちょうど２πの整数倍になるとき、上記第１実施形態で説明した膜厚解析方法における数５の式となる。上記第１実施形態で説明した膜厚解析方法ではψが整数倍変化する波長λ₁、λ₂を求めたが、λ₁、λ₂を任意に設定し、λ₁からλ₂の間の干渉光強度の波長変化（波形）から膜厚ｄを算出することも可能である。例えば、周波数解析を行うことにより、膜厚ｄを求める方法が採用できる。

When ψ is exactly an integral multiple of 2π, Equation 5 in the film thickness analysis method described in the first embodiment is obtained. Wavelength lambda ₁ above in the first embodiment in thickness analysis method described for ψ is an integral multiple changes have been sought lambda _2, lambda _1, lambda ₂ and the arbitrarily set, interference between lambda ₁ of lambda ₂ It is also possible to calculate the film thickness d from the wavelength change (waveform) of the light intensity. For example, a method of obtaining the film thickness d by performing frequency analysis can be employed.

Specifically, the interference signal having the wavelength as the horizontal axis has a waveform in which a subtle frequency modulation is applied to the center frequency. Therefore, ψ / 2π in Equation 18 can be replaced with a center frequency f having (λ ₁ −λ ₂ ) as a fundamental period, and can be expressed as in Equation 19. However, in the following formula, f has (λ ₁ −λ ₂ ) as a basic period (f = 1).

従って、この式からフーリエ変換を行うことで周波数ｆを算出し、この周波数ｆに基づいて膜厚ｄを算出することが可能となる。

Therefore, it is possible to calculate the frequency f by performing Fourier transform from this equation, and to calculate the film thickness d based on the frequency f.

  Furthermore, in the first embodiment, as a method for setting the analysis wavelength region, previously known numerical values (refractive index nox and film thickness dox) are input in advance to the analysis apparatus 5 and “numeric” is used using these numerical values. However, it may be as follows.

  That is, the maximum value is automatically detected from the waveform of the relationship between the wavelength and the reflection intensity obtained from the interference information from the spectroscope 4, and the wavelength region around the maximum value of the curve connecting the maximum values is set as the analysis wavelength region. You may make it do.

  In each of the above-described embodiments, the CCD array 4b is used as a means for collectively acquiring light interference information. However, other means such as a photodiode array may be used.

  In each of the above embodiments, even if only “nodes” are removed without using a means for obtaining interference information at once, such as a CCD array, the effect of removing “nodes” is sufficiently obtained. be able to.

  In the third embodiment, the case where the film thickness of the active layer 6a in the SOI substrate 6 is measured has been described. However, the correction shown in the third embodiment is used for the film thickness measurement of the two-layer film structure having different refractive indexes. It is also possible.

It is a figure which shows schematic structure of the film thickness measuring apparatus which implement | achieves the film thickness measuring method in 1st Embodiment of this invention. 2 is a diagram illustrating an example of a specific configuration of a spectroscope 4. FIG. It is a flowchart of the film thickness measurement using the film thickness measuring apparatus shown in FIG. It is a reference figure for demonstrating the film thickness analysis method. It is a wave form diagram which shows the relationship between a wavelength and reflection intensity. It is the figure which showed the light interference phenomenon typically. It is a wave form diagram which shows the relationship between a wavelength and reflection intensity. It is a wave form diagram which shows the relationship between a wavelength and reflection intensity. It is a wave form diagram which shows the relationship between a wavelength and reflection intensity. It is a wave form diagram which shows the relationship between a wavelength and reflection intensity. It is a wave form diagram which shows the relationship between a wavelength and reflection intensity. It is a wave form diagram which shows the relationship between a wavelength and reflection intensity. It is a wave form diagram which shows the relationship between a wavelength and reflection intensity. It is a wave form diagram which shows the relationship between a wavelength and reflection intensity. It is a figure which shows the wavelength dependence of the absorption coefficient of Si. It is the figure which showed the relationship between the wavelength which showed the thickness of the active layer 6a from which the transmittance | permeability becomes 10%, and the film thickness of the active layer 6a. It is a figure which shows the correlation of the measured value of a standard measuring device, and the measured value of 1st Embodiment. It is the figure which showed the measurement result at the time of rotating the SOI substrate 6. FIG. It is a wave form diagram which shows the relationship between the wavelength in the case of interposing water and reflection intensity. (A) is a wave form diagram which shows the relationship between the wavelength and reflection intensity when measuring to wavelength 1000nm, (b) is a figure which shows the relationship between the wavelength and reflection intensity in a theoretical value. It is the figure which showed schematic structure of the film thickness measuring apparatus in 2nd Embodiment of this invention. It is a flowchart of the film thickness measurement using the film thickness measuring apparatus shown in FIG. It is a figure for demonstrating the influence of a "section". It is a schematic diagram of the light interference phenomenon in the case of a two-layer film structure. It is the figure which calculated | required the relationship between the decimal part of a phase value, and the shift amount s by simulation. It is a figure which shows the actual measurement result which investigated the measurement error about the case where it corrects in 3rd Embodiment of this invention, and the case where it does not perform.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Light source 2 Optical fiber 3 Measuring head 4 Spectrometer 5 Analyzer 6 SOI substrate 6a Active layer 6b Support substrate 6c Oxide film

Claims (12)

Light is applied to the SOI substrate (6) formed by sandwiching the oxide film (6c) between the active layer (6a) formed of a semiconductor layer and the support substrate (6b), and reflected light from the SOI substrate. A film thickness measuring method for measuring the film thickness of the active layer based on
A step of setting an analysis wavelength region so as to exclude a wavelength at which light interference is attenuated by reflected light on both surfaces of the oxide film;
Spectroscopically reflecting the reflected light by irradiating the active layer with respect to each wavelength;
A step of collectively obtaining interference information of the divided light for each wavelength;
A relationship between the wavelength and the reflection intensity is obtained from the interference information in the analysis wavelength region, and a wavelength having a maximum value or a minimum value of a waveform showing this relationship is obtained, and the activity from the wavelength having the maximum value or the minimum value is obtained. And calculating the thickness of the layer,
In the step of calculating the thickness of the active layer, the wavelength at an arbitrary maximum or minimum value is λ ₁ , λ ₂ , the refractive index of the film at the wavelength λ is n (λ), and the frequency or wave number between both wavelengths is calculated. X, where s ₁ and s ₂ are the phase shift amounts at the wavelength of the maximum value or the minimum value in the single-layer film structure and the double-layer film structure, the film thickness d of the active layer is

The method for measuring a film thickness of a semiconductor layer, characterized by: 2. The phase shift amounts s ₁ and s ₂ are obtained based on a phase value of the oxide film obtained from a film thickness of the oxide film and a wavelength of light passing through the oxide film. The method for measuring the film thickness of the semiconductor layer according to the above. In the step of setting the analysis wavelength region,
Assuming that the refractive index of the oxide film is nox, the film thickness is dox, and m is an arbitrary positive number,

The method for measuring a film thickness of a semiconductor layer according to claim 1, wherein the film thickness is set in a range satisfying the above. 2. The step of calculating the thickness of the active layer calculates the thickness of the active layer from the wavelength that is the minimum value among the wavelengths that are the maximum value or the minimum value. 4. The method for measuring a film thickness of a semiconductor layer according to any one of 3 above. 5. The step of setting the analysis wavelength region, wherein a wavelength at which light transmittance in the active layer is a predetermined value or more is set as the analysis wavelength region. The method for measuring the film thickness of the semiconductor layer according to the above. 6. The semiconductor layer film according to claim 5, wherein in the step of setting the analysis wavelength region, a wavelength at which light transmittance in the active layer is 10% or more is set as the analysis wavelength region. Thickness measurement method. In the step of splitting the reflected light and the step of acquiring the interference information collectively, the diffraction grating (4a) for splitting the reflected light and the interference information for each wavelength of the light split by the diffraction grating are batched. A method for measuring a film thickness of a semiconductor layer according to claim 1, wherein a spectroscope (4) provided with a CCD array (4 b) obtained in this way is used. The method for measuring a semiconductor layer according to claim 7, wherein in the step of setting the analysis wavelength region, a wavelength of 1000 nm or less is set as the analysis wavelength region. 9. The semiconductor layer thickness measuring method according to claim 1, further comprising a polishing step of polishing the active layer with respect to the SOI substrate, wherein the thickness of the active layer during the polishing is determined by the method for measuring a thickness of the semiconductor layer according to claim 1. A method of manufacturing a semiconductor substrate, comprising: measuring and detecting that the active layer has a desired film thickness, and terminating the polishing. In the polishing step, the active layer is polished by rotating the SOI substrate, and the thickness of the active layer is measured while the SOI substrate is rotated. Item 10. A method for manufacturing a semiconductor substrate according to Item 9. Based on the reflected light from the substrate, the substrate (6) having the first layer (6a) and the second layer (6c) having a refractive index different from that of the first layer is irradiated. A film thickness measuring method for measuring a film thickness of the first layer,
Spectroscopically analyzing the reflected light by irradiating the first layer with respect to each wavelength;
Obtaining the interference information of the separated light for each wavelength and calculating the film thickness of the first layer using the acquired interference information,
In the step of calculating the film thickness of the first layer, the wavelength at an arbitrary maximum value or minimum value is λ ₁ , λ ₂ , the refractive index of the film at the wavelength λ is n (λ), the frequency between both wavelengths or When the wave number is X, and the phase shift amount at the wavelength that is the maximum value or the minimum value in the single-layer film structure and the two-layer film structure is s ₁ , s ₂ , the film thickness d of the first layer is

The method for measuring a film thickness of a semiconductor layer, characterized by: The phase shift amounts s ₁ and s ₂ are obtained based on the phase value of the second layer obtained from the film thickness of the second layer and the wavelength of light passing through the second layer. The method of measuring a film thickness of a semiconductor layer according to claim 11, wherein

Images